The ability to predict and retrodict the magnitude of urban heat islands (UHI) in the urban canopy layer (UCL) is of utility to those involved in "correcting" long-term temperature records, heating and air-conditioning, air pollution photochemistry, assessing urban mixing depths, etc. At present there is no operational scheme for the UHI which has the merits of simplicity and universality. One approach is to construct a purely statistical formulae similar to that pioneered by Sundborg, but the coefficients must be derived from pre-existing observations and are geographically specific to the city, and there is no temporal component. Another approach is to run a full numerical model but the input data requirements and expertise to run it are often too large for operational purposes and the output may not be applicable to the UCL.
An intermediate method is forwarded here, it derives from empirical evidence regarding the maximum magnitude of heat islands from many cities, their typical diurnal variation, and their modulation by weather controls. The maximum UHI for a given city on extensive flat terrain is known to occur at night in "ideal" (calm, cloudless) weather, and its magnitude is primarily related to urban-rural differences of horizon screening (sky view factor) and thermal properties (thermal admittance) of the substrate. Setting this as the absolute maximum value for a given city, increasing wind speed (a surrogate for turbulent mixing and advection), and increasing cloud amount / decreasing cloud height (surrogates for radiative exchange), are used to diminish the magnitude according to known empirical relations. Similarly, since the temporal variation of the UHI through a day can be shown to follow a broadly generalizable pattern, it is incorporated to provide first order temporal dynamics. Some element of universality is introduced by normalizing both the UHI magnitude - to that of the maximum for the city, and the temporal pattern - to the times of sunset and sunrise. The aim is to be able to retrodict UHI magnitude in the UCL for any city, in all weather conditions, and at any time of day within about one Celsius degree in the mean or up to about two degrees on an hourly basis. The predictive capability will degrade depending on the forecast accuracy.
At present the scheme is being tested for the case of a compact city in extensive flat terrain (uppsala) and new modules to accommodate intra-city site differences and the seasonal effects of anthropogenic heat flux, are being devised. The algorithms are not yet capable of handling cases where thermal advection dominates (synoptic or mesoscale fronts, cold air drainage) but in principle this should be possible.